RESUMEN
Objective:To understand the epidemiology, clinical characteristics and associated risk factors of severe adenovirus(ADV)pneumonia in children, providing the basis for targeted prevention and treatment.Methods:Clinical features of children with ADV pneumonia at Children′s Hospital of Soochow University from January 2011 to December 2020 were retrospectively analyzed.According to the severity of the disease, cases were divided into severe ADV pneumonia group and common ADV pneumonia group.The epidemiological and clinical characteristics of two groups were compared, and risk factors for the occurrence of severe ADV pneumonia were analyzed.Results:A total of 1 158 patients with ADV pneumonia were enrolled, including severe ADV pneumonia 104 cases(8.98%) and ordinary ADV pneumonia 1 054 cases(91.02%).The median age of severe ADV pneumonia group was 1.17 (0.83, 2.73) years, which was significantly younger than that of common ADV pneumonia group 3.16 (1.50, 4.50) years( P<0.05), and 77.89% (81/104) of them were younger than 3 years old.The occurrence of severe ADV pneumonia was predominant in winter and spring, accounting for 71.15% (74/104).Cough was present in 89.42% (93/104) and fever in 99.01% (103/104) of the severe ADV pneumonia group.Compared with the common ADV pneumonia group, the severe ADV pneumonia group had a significantly longer febrile time[10(6, 14)d vs. 5(4, 7)d, P<0.05], significantly higher incidence of shortness of breath, wheezing, convulsions/coma[100% vs. 2.09%, 45.19% vs. 13.57%, 10.57% vs. 1.99%, P<0.05], and significantly higher incidences of emphysema, pleural effusion, bronchial signs, pulmonary solids, and atelectasis [21.15% vs. 2.09%, 5.77% vs. 0.19%, 4.81% vs. 0, 3.85% vs. 0.09%, P<0.05].Multivariable Logistic regression showed that age younger than 1.71 years old, wheezing, and the presence of underlying diseases (moderate to severe anaemia, congenital heart disease, neurological disease) were risk factors for the development of severe ADV pneumonia ( P<0.05).Receiver operating characteristic curve analysis showed that the sensitivity and specificity of age<1.71 years old(20 months old) for predicting the occurrence of severe ADV pneumonia were 65.4% and 71.5%, respectively. Conclusion:The age of most severe ADV pneumonia is less 3 years in Suzhou.It usually occurres in winter and spring, with fever, cough, shortness of breath, and wheezing as the main symptoms.Pulmonary manifestations such as pleural effusion, emphysema, pulmonary consolidation, and atelectasis may occur.The underlying disease, wheezing, and age of onset less than 1.71 years (20 months) old are independent risk factors for severe ADV pneumonia.
RESUMEN
Artificial intelligence can potentially provide a substantial role in streamlining chest computed tomography (CT) diagnosis of COVID-19 patients. However, several critical hurdles have impeded the development of robust AI model, which include deficiency, isolation, and heterogeneity of CT data generated from diverse institutions. These bring about lack of generalization of AI model and therefore prevent it from applications in clinical practices. To overcome this, we proposed a federated learning-based Unified CT-COVID AI Diagnostic Initiative (UCADI, http://www.ai-ct-covid.team/), a decentralized architecture where the AI model is distributed to and executed at each host institution with the data sources or client ends for training and inferencing without sharing individual patient data. Specifically, we firstly developed an initial AI CT model based on data collected from three Tongji hospitals in Wuhan. After model evaluation, we found that the initial model can identify COVID from Tongji CT test data at near radiologist-level (97.5% sensitivity) but performed worse when it was tested on COVID cases from Wuhan Union Hospital (72% sensitivity), indicating a lack of model generalization. Next, we used the publicly available UCADI framework to build a federated model which integrated COVID CT cases from the Tongji hospitals and Wuhan Union hospital (WU) without transferring the WU data. The federated model not only performed similarly on Tongji test data but improved the detection sensitivity (98%) on WU test cases. The UCADI framework will allow participants worldwide to use and contribute to the model, to deliver a real-world, globally built and validated clinic CT-COVID AI tool. This effort directly supports the United Nations Sustainable Development Goals number 3, Good Health and Well-Being, and allows sharing and transferring of knowledge to fight this devastating disease around the world.